Amicus Curiae Brief — United States, Petitioner v. Andre Ralph Haymond

Supreme Court briefJan 25, 2019

Ask Donna

What actually matters in this document.

Text

No. 17-1672

IN THE

UNITED STATES OF AMERICA,

Petitioner,

ANDRE RALPH HAYMOND,

Respondent.

On Writ of Certiorari to the United States Court of

Appeals for the Tenth Circuit

BRIEF OF AMICUS CURIAE DUE PROCESS

INSTITUTE IN SUPPORT OF RESPONDENT

Shana-Tara O’Toole

DUE PROCESS INSTITUTE

700 Pennsylvania Ave., SE,

Suite 2019

Washington, DC 20003

David T. Goldberg

Counsel of Record

DONAHUE, GOLDBERG &

WEAVER, LLP

109 South 5th St., Suite 4201

Brooklyn, NY 11249

(212) 334-8813

david@donahuegoldberg.com

Table of Contents

Introduction and Summary of Argument ............... 1

The Determination Subsection (k) Calls

For Is One The Constitution Requires

Be Made By A Jury, Based On

Proof Beyond A Reasonable Doubt ................... 9

II. The Revocation Label Does Not Place

The Subsection (k) Regime Beyond

The Reach Of The Fifth and

Sixth Amendments .......................................... 19

III. The Government’s Remedial Proposal

Should be Rejected........................................... 32

Conclusion ............................................................. 34

Table of Authorities

CASES

Almendarez-Torres v. United States,

523 U.S. 224 (1998)...........................5, 12, 17, 30

Apprendi v. New Jersey,

530 U.S. 466 (2000)................................... passim

Barber v. Thomas, 560 U.S. 474 (2010)................. 20

Blakely v. Washington,

542 U.S. 296 (2004).................................... 10, 17

Bloom v. Illinois., 391 U.S. 194 (1968) ........ 5, 14, 15

Frank v. United States, 395 U.S. 147 (1969)......... 22

Gagnon v. Scarpelli, 411 U.S. 778 (1973) ...... passim

Gompers v. United States, 233 U.S. 604 (1914) 3, 15

Granderson v. United States, 511 U.S 39 (1994) .. 22

Green v. United States,

356 U.S. 165 (1958).................................. 1, 5, 15

Hudson v. Palmer, 468 U.S. 517 (1984) ................ 26

Int’l Union, UMW v. Bagwell,

512 U.S. 821 (1994)............................ 2, 4, 15, 17

Johnson v. United States,

529 U.S. 694 (2000)................................... passim

Jones v. Florida, 608 So.2d 797 (1992).................. 27

Knights v. United States,

534 U.S. 112 (2001).......................................... 26

Louisiana v. Brown,

645 So.2d 1282 (La. App. 4 Cir. 1994) ............ 27

Mempa v. Rhay, 389 U.S. 128 (1967) .................... 22

Mistretta v. United States,

488 U.S. 361 (1989).................................... 20, 27

Morrissey v. Brewer,

408 U.S. 471 (1972)................................... passim

Muniz v. Hoffman, 422 U.S. 454 (1975) ................ 16

In re Oliver, 333 U.S. 257 (1948) ........................... 18

Oregon v. Ice, 555 U.S. 160 (2009) ..........4, 11, 16, 18

Pennsylvania Bd. of Probation and Parole

v. Scott, 524 U.S. 357 (1998) ........................... 21

People v. Garbutt, 17 Mich. 9 (1868) ..................... 18

Ring v. Arizona, 536 U.S. 584 (2002) ................ 3, 10

Samson v. California,

547 U.S. 843 (2006).............................. 13, 25, 26

Schlup v. Delo, 513 U.S. 298 (1995) ...................... 13

Speiser v. Randall, 357 U.S. 513 (1958) ................ 13

iii

State ex rel. Bray v. Russell,

89 Ohio 3d 132 (2000) ...................................... 26

United States. v. Booker, 543 U.S. 220 (2005) .. 9, 13

United States v. Dixon,

509 U.S. 688 (1993)................................ 5, 15, 32

United States v. Grayson, 438 U.S. 41 (1978) ....... 20

United States v. Salerno, 481 U.S. 739 (1987) ...... 32

In re Winship, 397 U.S. 358 (1970).......................... 2

Wolff v. McDonnell, 418 U.S. 539(1974) ................ 26

Statutes

18 U.S.C. § 2252 ............................................. 2, 9, 12

18 U.S.C. § 3553(a)............................................. 2, 29

18 U.S.C. § 3559(a)................................................. 28

18 U.S.C. § 3559(a)(2)(A) ....................................... 28

18 U.S.C. § 3583(a)........................................... 14, 25

18 U.S.C. § 3583(c) ................................................. 29

18 U.S.C. § 3583(d)................................................. 14

18 U.S.C. § 3583(e) ................................................. 29

18 U.S.C. § 3583(k) ......................................... passim

18 U.S.C. § 3691 ..................................................... 16

S. Rep. No. 225, 98th Cong.,

1st Sess. 124 (1983) ..................................... 2, 25

OTHER MATERIALS

M. Frankel, Criminal Sentences:

Law Without Order (1973) .............................. 23

Fiona Doherty, Indeterminate Sentencing

Returns: The Invention of Supervised

Release, 88 N.Y.U. L. Rev. 958 (2013)....... 21, 28

Statement of Interest1

The Due Process Institute is

non-profit,

bipartisan, public-interest organization that works

nationwide to honor, preserve, and restore principles

of fairness in the criminal legal system. Formed in

2018, the Institute has already participated as

amicus curiae before this Court in a number of cases

raising important constitutional issues, such

as Timbs v. Indiana, No. 17-1091 (certiorari granted

June 18, 2018) and Turner v. United States, No. 18106 (petition for certiorari pending). Our concern

about the law before the Court echoes Justice Black’s

later-vindicated dissent in Green v. United States, 356

U.S. 165 (1958): “[E]xtraordinary authority [may]

first slip[] into the law as a very limited and

insignificant thing,” but unless promptly checked, it

may become a “pervasive mode of administering

criminal justice, usurping our regular constitutional

methods of trying those charged with offenses against

society.” Id. at 194.

Introduction and Summary of Argument

“This case involves basic questions of the highest

importance far transcending its particular facts.”

Green, 356 U.S. at 194 (Black, J., dissenting). The

regime the United States asks this Court to approve

is without precedent and contrary to our Nation’s

legal tradition. 18 U.S.C. § 3583(k) establishes a

proceeding where a person may be sentenced to life in

prison without parole, based on a single judge’s

finding that he “commit[ted] an[] offense” codified in

1 Pursuant to Rule 37.6, amicus affirms that no counsel for a

party authored this brief in whole or in part and that no person

other than amicus and counsel made a monetary contribution to

its preparation or submission. The Petitioner has consented to the

filing of this brief and the Respondent has filed a blanket waiver.

Title 18 of the U.S. Code. Even a close finding of

culpability by a preponderance of the evidence

suffices, though that standard has been held

constitutionally inadequate to impose eighteen

months’ juvenile detention, In re Winship, 397 U.S.

358 (1970), or a contempt fine, Int’l Union, UMW v.

Bagwell, 512 U.S. 821 (1994).

The sentence authorized is far longer than the

maximum Congress authorized for respondent’s 2010

conviction under 18 U.S.C. § 2252; it is far more

severe than the maximum Congress authorized for a

recidivist conviction. (Indeed, the Government’s

appears to take the position that a defendant

sentenced to life in a subsection (k) proceeding could

be prosecuted and punished additionally under

section 2252, Br. 51). And the subsection (k) penalty

is different in kind from the revocation sentences

provided for any other supervised release violation,

including the most serious. Had respondent

committed homicide while on supervised release

under the 2010 conviction, the maximum revocation

sentence would have been two years. And even then,

the district court would be under no mandate to order

revocation or any prison term. Rather, as with every

other violation of a supervised release condition, the

sentencing decision would be made in light of what

the Government agrees is the “predominant goal” of

supervised release, “to ease the defendant’s transition

into the community,” S. Rep. 98–225, at 125 (1984),

disregarding punishment; and in accord with the

“parsimony principle” Congress codified in 18 U.S.C.

§ 3553(a).

The Government seeks to obscure these stark

realities beneath a welter of labels, insisting countless

times

that

subsection

(k)

entails

merely

“administering a sentence,” rather than imposing

one; that the defendant is not being “prosecuted”

within the meaning of the Sixth Amendment; that a

court is merely “revoking” the defendant’s

“conditional liberty,” not convicting and punishing

him for the offense it has determined he committed.

I. None of that is tenable. As the Court has long

taught, when the substance of the Fifth and Sixth

Amendments collides with a label, the latter must

yield. “The fundamental meaning of the jury-trial

guarantee of the Sixth Amendment[—]that all facts

essential to imposition of the level of punishment that

the defendant receives” must be found beyond

reasonable doubt—does not vary, “whether the

statute calls them elements of the offense, sentencing

factors, or Mary Jane”—or “post-judgment sentencing

facts.” Ring v. Arizona, 536 U.S. 584, 610 (2002)

(Scalia, J., concurring).

No one looking at the text and structure of

subsection (k)—and no one wandering into Judge

Kern’s courtroom in 2016—would have any doubt that

he was adjudicating and imposing punishment for a

crime, an “infraction[] of the law, visited with

punishment as such.” Gompers v. United States, 233

U.S. 604, 610 (1914) (Holmes, J.); id. (“If such acts are

not criminal, we are in error as to the most

fundamental characteristic of crimes as that word has

been understood in English speech”). By the same

token, no observer would describe what was

transpiring as “sentence administration.” A judge

assigned to preside over a subsection (k) proceeding

needs to know next to nothing about the sentence she

ostensibly is “administering.” The gravity of the

conviction offense would be irrelevant, compare 18

U.S.C. § 3583(e)(3), as would the length of the

sentence of imprisonment and term of supervised

release imposed when the defendant was convicted, or

the maximum prison term that could have been

imposed. Indeed, Judge Kern, in imposing the

mandatory five-year sentence here, pronounced it

“repugnant” that subsection (k) required respondent’s

long imprisonment without a jury and proof beyond

reasonable doubt. Resp. Br. 2.

In our legal system, imprisonment is

presumptively punishment, and when a law provides

longer incarceration than could be imposed under the

criminal code for the same conduct—and the sentence

is life and the conduct is “commit[ting] an[] offense”—

the inference is irrefutable. It is fanciful to suggest

that subsection (k) rests on a congressional judgment

about “breaches of trust,” but if that were true,

defendants would still be entitled to a jury trial and

the protection of the beyond-reasonable-doubt proof

standard. See Int’l Union, UMW v. Bagwell, 512 U.S.

821 (1994).

The Government’s claims that it has combed the

historical record for a comparable jury trial right

betrays a certain obtuseness. Although the subsection

(k) regime is facially incompatible with the binding

interpretation of the Sixth Amendment established in

Apprendi v. New Jersey, 530 U.S. 466 (2000), it does

not take an Apprendi enthusiast to recognize that a

proceeding where a defendant stands to be punished

based on a determination of factual guilt or innocence

is within “the jury’s traditional domain.” Br. 31

(quoting Oregon v. Ice, 555 U.S. 160, 168 (2009)).

Rather, the burden is on the Government to identify

what, apart from a label—i.e., the fact that subsection

(k) appears in the supervised release provision of Title

18—distinguishes this proceeding from countless

others where a defendant’s imprisonment requires a

determination he “commit[ted] an offense under [a]

chapter [of Title 18].”

That respondent has “previously been found

‘guilty of a crime against the people,’” U.S. Br. 37

(quoting Morrissey v. Brewer, 408 U.S. 471, 480

(1972)) does not suffice. The Government cannot

contend that, in a prosecution under a repeat offender

statute, the fact of the defendant’s guilt on the second

offense may be determined without a jury. Cf.

Almendarez-Torres v. United States, 523 U.S. 224

(1998). Nor can it be that only persons with “absolute

liberty” are fully protected by the Fifth and Sixth

Amendments, see U.S. Br. 38; there is a long history

of affording the full panoply of rights to defendants

accused of committing crimes in prison, though they

occupy a far lower place on the “continuum” of liberty

than do persons on post-release supervision.

And, as this Court’s decisions concerning

punishment for contempt establish, it makes no

constitutional difference that the prohibition a

defendant is accused of violating appears in a

criminal statute, in a judicial decree, or (as here) in

both. See Bloom v. Illinois., 391 U.S. 194 (1968).

Indeed, to the extent that there is no long history of

juries determining alleged transgressions of “do not

violate the criminal law” orders, it is because courts

long refrained from issuing them, out of deference to

the “regular constitutional methods of trying those

charged with offenses against society.” Green, 356

U.S. 194 (Black, J., dissenting). But as that reticence

has receded, it has been settled that a defendant

accused of violating such an order is entitled to have

her guilt beyond a reasonable doubt be determined by

a jury. See United States v. Dixon, 509 U.S. 688, 695

(1993).

II. The Government’s defense ultimately rests on

a different label: “revocation.” If subsection (k) only

provided that a person (or even someone on

supervised release) committing an enumerated

offense would be sentenced up to life in prison, the

Government presumably would concede, its

codification in Section 3583 would not suffice to

extinguish defendants’ Fifth and Sixth Amendment

rights. Rather, the Government argues that because

it also provides for —automatic—“revocation” upon a

determination of guilt and because it references

Section 3583(e) (principally to say that section’s

limitations do not apply), a wholly different

constitutional analysis should govern, one that

permitted parole and probation revocations without

juries in Morrisey and Gagnon v. Scarpelli, 411 U.S.

778 (1973)—and that led the Court in Johnson v.

United States, 529 U.S. 694 (2000), to “appear[] to” see

supervised release revocations under subsection (e)(3)

as “unproblematic.” U.S. Br. 21.

That is wrong. In appearance, the modes of postrelease supervision may not have much changed

between the federal parole era and the present; nor

has the conduct of revocation hearings. But, as a

matter of constitutional substance, the world is

fundamentally different. The features that made

parole revocation hearings constitutional—that

defendants could not be ordered to serve more time in

prison than their sentence at trial imposed and

enjoyed “conditional liberty” that an early but

provisional release afforded—were swept away when

Congress decided in 1984 that time previously spent

on parole (two-thirds of a defendant’s sentence) would

have to be served in prison. And release at the end of

sentence no longer depends on the defendant’s

“promise” to comply. Br. 38. Release at the end of a

full sentence is a matter of entitlement, and the postrelease conditions are binding, with or without the

defendant’s agreement. Those conditions include

measures needed for effective supervision, including

searches that would be “unreasonable” under the

Fourth Amendment for others, but they don’t

establish a degraded citizenship status generally and

don’t extinguish defendants’ Fifth and Sixth

Amendment rights.

The Sentencing Reform Act also established a

brief and discretionary supervised release regime, as

a nonpunitive adjunct to supply the supportive

rehabilitative interventions that parole previously

provided and that some persons would still need on

first leaving prison. Congress subsequently added

provisions

explicitly

authorizing

sanctions—

including imprisonment for “brief periods”—in order

not to punish but to promote compliance and make

the supervision system operate more effectively.

That is the understanding of the law that the

Court adopted, at the Government’s urging, in

Johnson. That case did not raise a Fifth or Sixth

Amendment challenge to a prison sentence of any

kind, let alone to one like subsection (k), which was

not enacted until years later. But to the extent the

Court saw the regime before it as untroubling,

subsection (k) wipes away all the features that made

it so. Under the version of Section 3583 in effect in

Johnson, and still operative today for almost every

alleged supervised release violation, neither

revocation nor imprisonment was mandatory; and

periods of incarceration were subject to relatively

brief “caps,” which were and are determined by the

seriousness, not of the (alleged) violation, but rather

the offense of conviction. Indeed, under the version of

the law in effect when Johnson’s supervised release

was revoked, those caps applied to aggregates: The

most time respondent here could have spent in prison

if he repeatedly and seriously violated conditions was

two years in total. Those provisions were designed to

prevent—and do in fact prevent—what the

Government asks the Court to countenance here, a

“usurp[ation] of our regular constitutional methods of

trying those charged with offenses against society,”

whereby prosecutors, at their discretion, may obtain

severe criminal punishment under a much less

rigorous proof standard, based on the views of a

“single employee of the [Government]," Apprendi, 530

U.S. at 498 (Scalia, J. concurring).

III. The court of appeals here was correct—and

the Government is wrong—as to the appropriate

remedy. To the extent subsection (k) authorizes an

alternative procedure for finding a defendant guilty of

committing a federal offense and sending him to

prison, it is constitutionally impermissible, full stop.

That regime is inconsistent with the jury trial

guarantee

and

the

beyond-reasonable-doubt

requirement—and numerous other safeguards that

traditionally operate when a person is accused of an

offense that carries such serious punishment.

It makes no sense, as a matter of constitutional

law, congressional intent, or judicial administration,

for this Court to invent a different, hybrid procedure

for these cases, especially one that grafts some

constitutional protections, but not others, onto a

procedure designed and suited for making discrete,

highly discretionary, and non-punitive adjustments

to a defendant’s post-release status. The Tenth

Circuit’s straightforward resolution imposes no

special hardship: It allows the Government to address

allegations of repeat Section 2252 violations by

persons on supervised release in the same way it

treats fresh homicide allegations (indeed, the same

way it treats persons whose initial Section 2252

convictions pre-date the 2006 amendments to

subsection (k)). A defendant would be subject to

criminal trial for that offense and subject to

revocation and potential further incarceration, based

on a preponderance finding that his conduct violated

the “do not offend” condition that appeared in the

supervised release part of his initial sentence.

ARGUMENT

The Determination Subsection (k) Calls For Is

One The Constitution Requires Be Made By A

Jury, Based On Proof Beyond A Reasonable

Doubt

The repugnance Judge Kern expressed on

imposing sentence in accordance with subsection (k)

was warranted—and consistent with his oath of

office. Subsection (k) works a plain and serious

violation of the Constitution. As respondent carefully

demonstrates, the provision fails the Apprendi test in

stark fashion: subsection (k) requires defendants to be

sentenced to prison for at least five years—in addition

to the term that their conviction justified—based on a

“particular fact,” not found by the jury that rendered

a guilty verdict or any other, a fact that this law

“makes essential to [the] punishment” imposed.

United States. v. Booker, 543 U.S. 220, 232 (2005)

(quoting 542 U.S. at 301).

The Government does not seriously argue

otherwise, and its desultory efforts cannot succeed.

The Government posits that this prison term is

“authorized” by the original verdict, Br. 46—in the

sense that a conviction for Section 2252 automatically

triggers a term of supervised release, which in turn

requires the imposition of conditions and the

subsection (k) sentence in the event of a judiciallydetermined violation of an enumerated offense. But

that sort of argument is what Apprendi forecloses:

The defendant was on notice that a jury’s verdict

would result in twenty-year sentence if the judge

found a further aggravating fact by a preponderance

of the evidence, just as someone sentenced to the tenyear maximum for Section 2252 would be on notice

that he would spend at least fifteen years’

imprisonment if a judge found the “particular fact”

called for in subsection (k). The fatal Sixth

Amendment defect Apprendi identified applies fully

here: The judge could not, based on the jury findings

alone, impose the longer sentence. The “‘statutory

maximum’ for Apprendi purposes … [i]s not the

maximum sentence a judge may impose after finding

additional facts.” Blakely v. Washington. 542 U.S.

296, 303–04 (2004) (emphasis added). Accord Ring v.

Arizona, 536 U.S. 584, 602 (2002) (“[i]f a State makes

an increase in a defendant's authorized punishment

contingent on the finding of a fact, that fact—no

matter how the State labels it—must be found by a

jury beyond a reasonable doubt”).

The Government’s principal argument is not that

subsection (k) satisfies the Sixth Amendment test

Apprendi and its progeny establish, but rather that it

is exempt from it. The Apprendi line, the Government

insists, does not require jury determination of “postjudgment sentencing facts,” Br. 30—only facts that

could be alleged in an indictment. That is a doubtful

description of the Apprendi doctrine, and it is a wholly

untenable account of the Sixth Amendment. Neither

Apprendi nor its progeny is fairly read as announcing

or enforcing the limit the Government describes.

Those opinions said “any” and “all” facts, and all but

one granted relief to defendants, on the ground that

the legislature’s designation of an aggravating fact as

a “sentencing factor,” rather than an element of the

offense, could not trump the accused’s constitutional

right to have punishment limited to the maximum the

jury verdict alone allowed. And the lone decision

rejecting an Apprendi claim, Oregon v. Ice, 555 U.S.

160 (2009), hardly supports the Government here.

The defendant in that case received two separate

sentences, each supported by a jury determination of

guilt, and the Court held that the Sixth Amendment

did not require jury involvement in deciding whether

they should be served consecutively; respondent here

received sentences based on one jury verdict and one

judicial finding.

The Sixth Amendment problem here is much

more basic than what has confronted (and often

divided) the Court in Apprendi and progeny. The

“fact” on which subsection (k) defendants are denied

jury determination—and that subjects them to a fiveyear-to life prison sentence is one that has never,

under any sentencing regime, been considered a

“sentencing factor” or otherwise committed to judge,

rather than jury, resolution: whether or not he is

guilty of “commit[ting] an offense” established under

a chapter of the federal criminal code.

The precedent for having a jury determine that

question is not scant or hard to find; it is ubiquitous—

it is what happens—indeed, what the Constitution

demands—in every criminal prosecution under

Section 2252 and for every other non-petty offense.

That the facts alleged to state an offense under

chapter 110 in this and other subsection (k)

proceedings are “post-judgment,” Br. 20, and thus

entirely distinct from ones on which the jury rendered

an earlier guilty verdict is not argument against jury

determination. This Court has considered what the

Sixth Amendment requires in cases, like those under

subsection (k), where the punishment a defendant

faces is based on the fact that he has “already ‘been

convicted of a crime,’” U.S. Br. 18 (quoting Gagnon).

In Almendarez-Torres v. United States, 523 U.S. 224

(1988), the Court reaffirmed that it is constitutional

and appropriate to treat repeat offenses as distinct,

aggravated crimes, and further held, over forceful

dissent, that when a defendant is accused of being a

repeat offender, that fact need not be proven to a jury.

And so it is where a person previously convicted of

Section 2252 is accused for a second time of

committing that offense. But nowhere there, or

anywhere else, has it been suggested that in such

cases, the determinations of factual guilt for both

offenses requires anything but jury determination,

under the beyond-a-reasonable-doubt standard.

(Indeed, there is no question that the many other

safeguards—including thee rights to a trial in the

State and district where the offense occurred, to

confront witnesses, and to compulsory process—are

also guaranteed). Compare Pet. App. 31a-32a (Kelly,

J., dissenting) (arguing that it was sufficient that

respondent “was tried and found guilty beyond a

reasonable doubt of the original offense” and received

“the full panoply of rights during his initial criminal

proceeding”).

Thus, it cannot be denied that a prior conviction

for a similar offense does not in itself effect a life-long

forfeiture of the Constitution’s safeguards against an

unfair trial or an unjust conviction arising from “postjudgment” facts. Nor does a conviction effect even a

temporary forfeiture. The Government surely does

not dispute that if a person just beginning to serve a

sentence is accused of committing a similar offense in

prison is entitled to the “full panoply” of trial

protections, see, e.g., Schlup v. Delo, 513 U.S. 298, 304

(1995), even though he occupies an unfavorable place

on the “‘continuum’ of state-imposed punishments.”

Samson v. California, 547 U.S. 843, 850 (2006).

Indeed, it is doubtful that the Constitution would

allow a defendant to voluntarily surrender Fifth and

Sixth Amendment rights in future proceedings, in

exchange for the valuable benefit of leniency on a

pending charge. Cf. Speiser v. Randall, 357 U.S. 513

(1958). But subsection (k) does worse than that: It

imposes that forfeiture, in many cases for life, on

defendants who have no choice in the matter and who

have received and served the full punishment for

their conviction—including, in some cases, the

maximum Congress permitted to be imposed on

anyone convicted of the offense. It is incumbent on the

Government to articulate a “distinction of

constitutional significance,” United States v. Booker,

543 U.S. 220, 233 (2005), between accused second

offenders (1) who either are wholly free or still in

prison, and (2) those who are on supervised release.

The fact that the latter group are, by virtue of the

supervised release “part of the[ir] sentence,” under

court order to not violate Section 2252 does not supply

a constitutionally sufficient distinction. On its own

terms, the premise that subsection (k) imposes

punishment for a “breach of trust,” and not for

committing

the

offense,

warrants

maximal

skepticism. The penalty authorized is a mandatory

prison sentence with a longer maximum than the

recidivist Section 2252 criminal offense carries. While

no one expects Congress to make “every possible

supervised-release violation result in reimprisonment

for the exact same length of time,” U.S. Br. 49

(emphasis added), it is hard to credit that a

legislature willing to authorize this punishment

would categorically forbid district courts from more

than two-year sentences for disobeying conditions the

court itself imposed or for breaking the “promise” to

not commit homicide. Br. 38.

The Government points to no evidence that

Congress

actually

made

the

“categorical

determination” Judge Kelly hypothesized. The

condition alleged to have been violated did not derive

from the court whose “trust” respondent was alleged

to breach, but rather through operation of three

statutes Congress enacted, see 3583(a) (requiring that

persons convicted of sexual offenses serve supervised

release); id. § 3583(d) (requiring that district courts

prohibit); id. § 2252 (prohibiting knowing possession

of pornographic depictions of minors), and the district

court made clear that the breach did not warrant

more than two years’ imprisonment, Resp. Br. 8, and

that it would have acquitted had the criminal proof

standard governed. Pet. App. 68a.

But if disobedience were the actual basis for

imposing punishment, that would not make the

regime constitutional. This Court’s decision a half

century ago in Bloom v. Illinois, settled that the

Constitution does require jury trials and proof beyond

a reasonable doubt in cases where disobedience of a

court’s order is the basis for imposing punishment.

That decision disavowed contrary statements in more

than a century of opinions in favor of the logic

articulated in Justice Holmes’s Gompers opinion and

Justice Black’s Green dissent: that judicial “decrees

are simply another form of sovereign directive aimed

at guiding the citizen’s activity” and do not occupy

“any higher or different plane than the laws of

Congress … insofar as punishment for their violation

is concerned.” 356 U.S. at 219.

These cases refute the Government’s claims that

“post-judgment” jury fact-finding is unheard of or

historically anomalous. Indeed, supervised release

orders are essentially injunctions, which impose

affirmative and prohibitory obligations, not least to

refrain from violating criminal laws. Notably, that

sort of directive, which now appears in almost every

federal criminal judgment—is of recent origin. For

centuries, equity courts refused to enjoin crimes,

because doing so would necessarily impinge on the

regular constitutional means for enforcing criminal

laws. See United States v. Dixon, 509 U.S. 688 (1993).

But in modern practice, alleged violations of those

orders are held trigger the same full protections. See

id. at 695; see also Bagwell, 512 U.S. at 843 (Scalia,

J., concurring) (“As the scope of injunctions has

expanded, they have lost some of the distinctive

features that made enforcement through civil process

acceptable.”).

In fact, even before this Court reversed course in

Bloom, Congress by statute provided (for cases where

the United States is not a party) that persons accused

of a contempt that “also constitutes a criminal offense

under any Act of Congress, or under the laws of any

state [are] … entitled to trial by a jury, which shall

conform as near as may be to the practice in other

criminal cases.” 18 U.S.C. § 3691, and further that

imprisonment in such cases may not “exceed the term

of six months.” id. § 402. Cf. Muniz v. Hoffman, 422

U.S. 454, 475 (1975) (jury trial not constitutionally

required for criminal contempt if punishment is less

than six months). This 70-year-old federal regime

adopts exactly the distinction the Government here

brushes aside as making “little sense.” Br. 48-49.

Indeed, a further reason for recognizing jury trial

rights in contempt cases has special resonance in the

subsection (k) setting. As the Court emphasized in

Blakeley, the Sixth Amendment right is a

“fundamental

reservation

of power

our

constitutional structure,” and meant to “control the

judiciary.” 542 U.S. at 305-06. Accord Apprendi, 530

U.S. at 547-48 (O’Connor, J., dissenting) (the jury

trial guarantee functions “to protect the criminal

defendant against potentially arbitrary judges”). In

cases like this one, as with contempt, it is a

constitutional serious concern, not a virtue, that the

person initiating the proceeding—and authorized to

determine guilt and punishment—is often the same

one whose trust defendant allegedly breached and

who previously earlier adjudged the defendant guilty.

Cf. U.S. Br. 31 (suggesting that the fact revocation “is

ordinarily initiated by a judicial officer or the court

itself, not by a prosecutor” is reason for withholding

jury trial). Because judges “remain human even after

assuming their judicial duties,” and are not immune

to “pettiness and bruised feelings,” subsection (k)

defendants need the “historic … bulwark” juries

provide at least as much as others do. U.S. Br. 19

(quoting Ice, 555 U.S. at 168).

The record in this case raises no immediate

concern on that score. The district court appears to

have been fair and admirably forthright in its

dealings with respondent (though it did commit clear

error in its factual finding against him). But the

power subsection (k) vests in a single individual—to

impose not six months’ imprisonment or two years’,

but rather life without parole, based on his own view

of the facts, is wholly extraordinary.2

The “historical[] anomal[y]” in this case, U.S. Br.

20, is that a federal statute authorizes life

imprisonment without parole based on a judge’s belief

that a defendant, more likely than not, committed a

crime. Subsection (k) replicates the hypothetical

regime that all—“even Apprendi’s critics”—“would

reject” as a constitutionally intolerable, “absurd

[Sixth Amendment] result,” Blakely, 542 U.S. at 306:

It authorizes the second-most-serious punishment

known to the law based on a verdict of guilt for an

unrelated Class C felony and a judicial fact-finding.

And it implicates the core purposes of the jury-trial

2This is not to suggest that the precise regime governing

contempt trials should—or must—be transplanted to supervised

release revocation hearings or that defendants are entitled to

juries in every proceeding where incarceration is theoretically

possible. Many revocation sentences are brief and aimed at

securing compliance with conditions, and would thus not trigger

jury trial rights if the contempt analogy governed. See Bagwell,

512 U.S. at 829. And in many cases where a violation of the “do

not re-offend” command is the basis for seeking revocation, the

defendant has already been criminally convicted, in which case

the Almendarez-Torres rule would foreclose any jury plea. See,

e.g., Johnson, 529 U.S. 697. And it surely is constitutionally

significant that the avowed object of the regime in which such

proceedings occur, unlike that of civil trials, is to benefit the

defendant—and that judges’ powers are constrained by statute

in ways that have no analogue in contempt law.

guarantee and the core competence of the AngloAmerican criminal jury:

[F]rom its numbers, the mode of their

selection, and the fact that jurors come from all

classes of society, [the jury is] better calculated to

judge of motives, weigh probabilities, and take

what may be called a common sense view of a set

of circumstances, involving both act and intent,

than any single man, however pure, wise and

eminent he may be.

People v. Garbutt, 17 Mich. 9, 27 (1868) (Cooley, J.)).

There can be no suggestion here, very much

unlike with other claims considered in the Court’s

recent decisions, that the Sixth Amendment’s

purposes—or defendants’ interests—point in both

directions. See, e.g., Oregon v. Ice, 555 U.S. 160, 171

(2009) (requiring jury involvement in structured

process for making concurrent sentencing decision

could reward reversion to wholly discretionary

judicial decisionmaking); Apprendi, 530 U.S. at 546

(O’Connor, J., dissenting) (objecting that Court’s rule

affords preferential Sixth Amendment treatment to

entirely discretionary judicial sentencing). And there

countervailing

governmental

interest

comparable to the historic “broad power to define

crimes and their punishments,” id. at 524—only

boiler-plate assertions of practical difficulty. But

“restrictions upon authority for securing personal

liberty, as well as fairness in trial to deprive one of it,

are always inconvenient—to the authority so

restricted.” In re Oliver, 333 U.S. 257, 280-81 (1948)

(Rutledge, J., concurring).

II. The Revocation Label Does Not Place The

Subsection (k) Regime Beyond The Reach Of The

Fifth and Sixth Amendments

The

Government

has

identified

constitutionally significant basis for affording

drastically different procedures for adjudicating guilt

and imposing punishment within the group of

defendants who, having previously been convicted of

a sexual offense, are accused of committing another.

When the Government nonetheless announces

that “history, tradition, constitutional text” all prove

the constitutionality of Section 3583(k), its claim is

limited to a label-based argument: (1) that Morrissey

and Gagnon held that parole and probation may be

revoked based on a judicial determination of a

violation; (2) that supervised release revocation is

close enough to parole revocation; and (3) that

subsection (k) proceedings, in turn, are similar

enough to ordinary revocation hearings to keep the

Fifth and Sixth Amendments at bay. Indeed, the

often-exotic verbal formulations in the Government’s

brief appear to have been reverse-engineered to

reinforce this syllogism: Prison sentences are

“deprivations of conditional liberty”; findings of guilt

and innocence are “sentence implementation facts”;

etc.

But the Government’s second and third premises

are both wrong. Although the present-day supervised

release regime is, in appearance and daily operation,

a continuation of earlier mechanisms—imposing

restrictions, requirements, and supports to enable

successful reintegration—it also operates in a larger

framework that is, in almost every way, a deliberate

repudiation of its predecessor. Those fundamental

differences make supervised release revocation

diverge in important and constitutionally significant

respects from parole revocation. The features that

made the regimes in Morrissey and Gagnon

essentially immune from Fifth and Sixth Amendment

invalidation and maintained a sharp division between

“sentencing implementation” and criminal law

enforcement no longer operate. Instead, the

Sentencing Reform Act relies on a suite of statutory

provisions that together ensure that revocation and

incarceration remain mechanisms ancillary to

integrating persons leaving prison and not a means of

“usurp[ing] our regular constitutional methods of

trying those charged with offenses against society.”

Those features played a central role in the

Government’s defense of the supervised release

regime in Johnson. It is precisely those safeguards

that subsection (k) jettisons.

First, while parole was the “precursor” of

supervised release as a monitoring regime, U.S. Br.

21, it is error to say that Congress replaced parole

with supervised release. Rather, the Sentencing

Reform Act replaced parole with incarceration. In the

pre-1984 era, federal sentencing was indeterminate;

courts, “as a general rule, [could] conditionally release

a prisoner any time after he serve[d] one-third of the

judicially fixed term,” United States v. Grayson, 438

U.S. 41, 47 (1978); and many defendants in fact

served no more than that. See Barber v. Thomas, 560

U.S. 474, 482 (2010). The Sentencing Reform Act

brought an end to that, “mak[ing] all sentences

basically determinate,” in the interest of “uniformity

and greater honesty.”: “The sentence the judge

imposes [became] the sentence the offender actually

serve[s].” Id. (quoting Mistretta v. United States, 488

U.S. 361, 367 (1989)). The supervised release regime

was in fact a late addition, a recognition that though

Congress meant a decisive break from the

individualized, rehabilitative focus of the prior

regime, there would still be defendants leaving prison

after serving full-term sentences who might benefit

from the kind of support and monitoring formerly

extended through the parole process. See generally

Fiona Doherty, Indeterminate Sentencing Returns:

The Invention of Supervised Release, 88 N.Y.U. L.

Rev. 958 (2013).

In contrast, the “essence” of the former parole

system was “release from prison, before the

completion of sentence, on the condition that the

prisoner abide by certain rules during the balance of

the sentence.” Morrissey, 408 U.S. at 477 (emphasis

added); Pennsylvania Bd. of Probation and Parole v.

Scott, 524 U.S. 357, 365 (1998) (“In most cases, the

State is willing to extend parole only because it is able

to condition it upon compliance with certain

requirements.”). And that early release could be

recalled or revoked, and the “prisoner,” retaken, in

the event of noncompliance, at which point the

remainder of the sentence would be reinstated,

subject to future parole opportunities. Under that

system, what happened in respondent’s case—a

defendant’s being subject to five years’ imprisonment

beyond the term he received on conviction—was

inconceivable, indeed logically impossible, let alone

serving more time in prison than the statutory

maximum for a conviction offense, let alone life

without parole.

In a real sense, no Sixth Amendment problem was

possible under this system. Upon revocation, a

defendant served the unserved portion of his sentence

in prison (or out, if subsequently paroled), and he was

released. The defendant’s overall period of

involvement might exceed the length of his sentence,

and the time spent in prison, greater than if he had

never violated the terms of parole, but no defendant

served more prison time than the verdict of conviction

allowed.3

Probation was (and is) essentially identical,

particularly in cases where the sentence was imposed

upon conviction, but its execution is suspended.

Probation was generally thought even more

advantageous to the defendant than parole, because

he might never set foot in prison (rather than serving

one-third of a sentence). But like parole revocation,

and unlike revocations in the present-day

“determinate” regime, the outer limit of a revocation

proceeding was reinstatement of the sentence

imposed at the time of conviction—with the

possibility of parole (and no further supervision once

that sentence was served).4

3 To be sure, defendants whose parole was revoked near the

end of their term were not entitled to credit for the time they

were “on the street,” subject to reporting obligations and with

less than complete freedom. But defendants whose supervised

release is revoked lose that, too, in addition to extra time they

serve in prison (and under post-release oversight). It has in any

event been this Court’s teaching that imprisonment is a uniquely

serious form of punishment and that the exchange rate with

noncustodial supervision regimes is very steep. See Granderson

v. United States, 511 U.S 39 (1994); Frank v. United States, 395

U.S. 147 (1969).

4 Ordering probation by suspending imposition of sentence

was somewhat less defendant-friendly; the defendant still

emerged from sentencing with the prospect of never going to

prison, but any sentence he did receive would be imposed at a

juncture when the court had found a violation warranting

revocation. That, presumably, is why the Court in Mempa v.

Rhay, 389 U.S. 128 (1967), treated such proceedings differently

than the ones in Gagnon, notwithstanding the “post-judgment”

There were many serious objections to that

sentencing regime, including to the administration of

revocation hearings.5 But revocation proceedings of

the kind before the Court in Morrissey and Gagnon

were per se compliant with Apprendi. Moreover,

without minimizing the individual hardship of being

returned to prison, it is not hard to see how many

courts saw parole revocation as more a breach-ofcontract remedy rather than a punishment.

Defendants had an option to reject the restrictions

early release entailed, by not seeking parole; the time

remaining on the original sentence, and not the new

violation defined the revocation “sentence”; and,

consistently with the regime’s rehabilitative focus,

reimprisonment was itself temporary, given future

parole opportunities. When this Court and others

described parole revocation as a deprivation of

“conditional liberty,” Morrissey, 408 U.S. at 483, they

were not speaking in a loose or metaphorical way: A

defendant achieved absolute liberty once he had

completed serving his full sentence; “conditional

liberty” referred to the time spent outside prison, by

virtue of a paroling authorities’ decision to release

him early.

posture. But those defendants could not be required to serve

more than the sentence imposed, which could not exceed the

maximum authorized.

5 Among the complaints leveled at this regime were that

initial sentencing inscrutable, unpredictable, and nonuniform;

that parole authorities exercised releasing discretion arbitrarily

and discriminatorily and likewise treated those under

supervision unequally. See, e.g., M. Frankel, Criminal

Sentences: Law Without Order (1973).

Persons subject to supervised release under

present law are not given the benefit of early release.

Release is nondiscretionary; the defendant must have

served the entire prison sentence his conviction was

determined to warrant (minus good time credit), at

which point 18 U.S.C. § 3624(a) mandates he be

released (or on the proceeding Friday, if the term ends

on a weekend). Persons on supervised release are not,

as the Government implies, “allow[ed],” U.S. Br. 5, to

serve “part of,” id.¸ their prison terms at liberty, “on a

promise” to not violate conditions, id. 38.

Respondent’s conditions were imposed, not accepted

and would have been no less binding had he not

“signed th[em]…and acknowledged the consequences

for violating them.” Id. 9. Whereas under the parole

system, no person revoked could serve more time in

prison than the term originally imposed, now every

person who is incarcerated as a consequence of a

supervised release violation is serving time in

addition to his prison term.

Nor, despite the Government’s mantra-like

repetition of the phrase, does the fact that the liberty

of persons subject to subsection (k) is limited, mean

that it is “conditional” in the sense Morrissey used

that term. When a court “revokes” a defendant’s

supervised release and sentences him to

imprisonment, it is not rescinding or calling back an

earlier grant of liberty nor reinstating an unserved

portion of the prison sentence imposed on the jury’s

verdict. Persons on supervised release have

“conditional liberty” only in the sense that others in

free society does: They enjoy freedom, subject to the

condition that it may be taken away if they violate a

statute or court order.

The Government is of course correct that persons

subject to supervised release, by virtue of that status,

do not enjoy the full measure of liberties during the

time that “part of the[ir] sentence” operates. Br. 5,

(quoting 18 U.S.C. § 3583(a)). But their liberty is not

something that has been granted, and the limitations

are imposed, for non-punitive reasons, upon

conviction, in order to improve the defendant’s

prospects for reintegration, largely in recognition of

the legal and social disabilities that those leaving

lengthy incarceration experience. S. Rep. No. 225,

98th Cong., 1st Sess. 124 (1983) (a “primary goal [of

supervised release] is to ease the defendant’s

transition into the community after the service of a

long prison term for a particularly serious offense”).

Accordingly, the kinds of restrictions that are

commonly imposed post-release that could not be on

someone at absolute liberty are ones inherent in

advancing the regime’s supervisory goals. It is

“reasonable,” this Court has held, see Samson, v.

California, 547 U.S. 843 (2006), that persons under

rehabilitative supervision are searched more readily

than those who have regained absolute liberty. But it

does not follow—and it is incorrect—that the Fifth

and Sixth Amendment rights to a fair and accurate

determination of guilt for a different alleged “offense

under chapter 109A…” are likewise degraded. Indeed,

even with respect to the Fourth Amendment, it is

unlikely a police officer’s using what would normally

be unconstitutionally excessive force would be made

reasonable if the arrestee were on federal supervised

release.

In fact, the very decisions the Government

invokes in support of its status-based theory expose

the error of that argument. This rule applied in

Samson and Knights v. United States, 534 U.S. 112,

(2001), was derived from precedent establishing that

prisoners have no reasonable privacy expectations,

Hudson v. Palmer, 468 U.S. 517, 527 (1984); parolees’

and probationers’ place in the middle of the

“continuum” of liberty were held to support

correspondingly intermediate privacy rights. But, as

explained above, the same class denied Fourth

Amendment protection altogether, are understood,

universally, to enjoy the full panoply of fair-trial

rights if accused of committing crime while serving an

incarceration term. (The one seeming exception—

prison disciplinary proceedings—is not; as with

parole, what is at stake is not additional punishment,

but instead withdrawal of previously bestowed “good

time credits,” see Wolff v. McDonnell, 418 U.S.

539(1974); cf. State ex rel. Bray v. Russell, 89 Ohio 3d

132 (2000) (invalidating “bad-time” credit statute on

state constitutional grounds)). The Constitution is not

reasonably construed as requiring that all trial

safeguards be accorded a defendant charged with a

brutal killing while serving his homicide sentence in

prison, but countenancing subsection (k), which

authorizes life imprisonment if a judge finds it more

likely than not that a defendant who completed a

prison sentence committed a Class C felony.6

6 On the Government’s theory, Congress could provide for a

death sentence to be imposed based on a judge’s finding of guilt

for homicide at a “revocation” hearing for a defendant on

supervised release for a Class E felony. That, too, would not be

punishment or a “Prosecution,” but a mere “administration of a

previously imposed sentence” based on “post-judgment facts.”

The unlikelihood the Government would be “[]willing[]to follow

its premise to that logical conclusion suggests that the premise

itself is flawed.” U.S. Br.47.

The Government at points floats the notion that

the present system is not merely like prior regimes;

but in fact is, in substance, a form of probation—a

“split” sentence, where the defendant serves a first

part in prison and a second “part” is suspended and

served, “provisionally,” Br. 5, outside. That is not the

regime the Congress enacted or the one the

Government previously described to this Court, and it

would raise far-ranging problems in the vast run of

federal criminal cases where a supervised release

term is imposed. It would be startling enough that a

statute enacted to “make[] all sentences basically

determinate,” Mistretta, 488 U.S. at 367, in fact

established a regime where almost every convicted

defendant receives a paradigmatic indeterminate

sentence. But more important, treating the

supervised release “part” of the sentence as if it were

a suspended prison term would render unlawful

sentences imposed routinely in the federal system,

including the one here: In jurisdictions that permit

split sentences, a sentence of three years in prison

plus an additional ten years suspended is

impermissible if the statutory maximum sentence is

ten. See, e.g., Jones v. Florida, 608 So.2d 797, 800

(1992); Louisiana v. Brown, 645 So.2d 1282 (La. App.

4 Cir. 1994).

But as the Government explained to the Court in

Johnson, the supervised release term, is not imposed

as punishment. Rather, its “dominant purpose” is “to

afford offenders assistance in reintegration into

society and rehabilitation.” Br. U.S., No. 99-5153, at

15. A court’s authority to revoke and further

incarcerate is not central to the regime; but rather an

incident Congress included to better accomplish the

regime’s nonpunitive purposes. Id. 27. Indeed, the

fact that initial versions of the Sentencing Reform Act

omitted any mechanism for enforcing compliance

with conditions itself attests that Congress saw

supervision as an end in itself and refutes the

suggestion that it had re-enacted parole. See Doherty,

88 N.Y.U. L.J. at 999.

Since Congress enacted what is now subsection

(e)(3), providing further imprisonment as an available

response to violations, it has taken care to ensure that

revocation proceedings not become an alternative

mechanism for punishing violations of criminal law

that happen to occur while the person is under

supervision. First, Congress enacted the limitations

in subsection (e)(3) of the statute, which establish

maximum terms of incarceration of one, two or three,

or five years, depending on the severity of the

defendant’s offense of conviction. These ceilings,

which are a small fraction of the sentences provided

for the corresponding offenses, see 18 U.S.C.

§ 3559(a), signal that incarceration is not meant to

be—and cannot supply—punishment for a serious or

recidivist offense. And these finite and relatively low

ceilings are suited to the informal and holistic

character of the proceeding at which sanctions are

decided.

That signal is reinforced by the fact that the

graduated maximums are linked to the seriousness of

the conviction offense, not the conduct underlying the

violation. And it is further reaffirmed by the broad

discretion courts are granted: they need not revoke in

response to every proven violation, or impose the

maximum when they do decide to order incarceration.

Congress conspicuously omitted “the need for just

punishment,” 18 U.S.C. § 3559(a)(2)(A), from the list

of considerations to be taken into account when

imposing supervised release and when deciding

whether and how to respond to violations—on the

understanding that the defendant’s original prison

term would have fully accounted for that. See 18

U.S.C. § 3583(c), (e). Congress further directed courts

to take account of the individual’s educational,

medical, and vocational needs and “impose a sentence

sufficient, [that is] but not greater than necessary.”

id. § 3553(a)(2)(D), (a).

Finally, as the Government emphasized in

Johnson, the power to sanction may not be

understood in isolation from “the other powers

granted to the district court in Section 3583(e), all of

which are designed to ensure a successful transition

for the defendant from prison to society outside

prison…The district court [is] granted a continuing

supervisory authority, during the entire period of

supervised release, to monitor the progress of the

offender and to adjust the conditions of supervised

release to assure his successful reentry into the

community.” Johnson Br. 27.

That is largely how the regime operated when this

Court decided Johnson. And that is how it operates

with respect to every defendant on supervised release

except those in respondent’s position. Subsection (k),

of course, does none of this. It strips district courts of

all discretion, mandating that revocation occur,

without regard to considerations a court is otherwise

required to consider, 18 U.S.C. § 3583(e), and a

lengthy prison sentence be imposed—not “a short

period of reimprisonment, followed by a new

opportunity … to achieve successful reintegration,”

Johnson Br. 31. Life sentences aside, the subsection

(k) minimum equals the maximum that may be

imposed on any defendant convicted of Class A

felonies who commits a second (or third) Class A

felony on supervised release. And punishment is tied

exclusively to the “offense” the defendant is alleged to

have committed while on supervised release.

In light of this, it is unsurprising that the

Government does not echo the dissenting opinion

below in claiming that this Court’s Johnson decision

“already” acquitted the subsection (k) regime of

violating the Sixth Amendment. See Pet. App. 34a.

Neither subsection (k), which was enacted in 2006,

nor any provision like it was before the Court when

Johnson was decided. Nor was any provision of

Section 3583(e) challenged on Sixth Amendment

grounds. (Because the principal ground for revocation

was Johnson’s recent state court conviction, 529 U.S.

at 697, no plausible Sixth Amendment challenge

could have been raised, see Almendarez-Torres).

Indeed, the petitioner raised no challenge,

constitutional or otherwise, to the incarceration

portion of his revocation sentence. The only issue

before

the

Court

concerned

court-imposed

supervision: whether a district court was prohibited

from imposing further supervision if it found a

violation and sent the defendant to prison. This Court

resolved that question on statutory grounds,

reasoning that the text of subsection (e)(3) did not

require that reading, i.e., that “revoking” supervised

release status did not necessarily mean “terminating”

it, 529 U.S. at 704, noting that its reading comported

with the purposes and design of the supervised

release regime, id. at 708-10.

The Court did not answer the one constitutional

question that was raised: Whether, for Ex Post Facto

Clause purposes, a defendant-unfriendly amendment

to the supervised release regime should apply to only

those whose sentences are imposed thereafter or

whether it may also apply to those whose violation

conduct post-dates the amendment. The answer the

Johnson signaled, that the sanction should be

“attribute[d],” 529 U.S. at 701, to the conviction, was

the position that both parties urged the Court to

adopt, and one of the grounds was that doing so would

avoid “serious constitutional difficulties,” id. at 700.

The Government’s inference from that discussion

and—that the Court viewed the supervised release

revocation regime, including the limited power to

incarcerate, as “unproblematic,” Br. 21—is not

implausible or incorrect. The sanction regime then in

effect was more modest than the one that now

governs—sentences were limited by both the offensebased ceilings and by “the term of supervised release,”

imposed on the individual defendant, id. at 703, and

the subsection(e)(3) ceilings aggregated all revocation

imprisonments; the Government took pains to tell the

Court its “view… [that] an offender like [Johnson],

who was convicted of a Class D felony faces only a

maximum of two years’ reimprisonment for all …

violations of supervised release,” explaining “[a]t

some point the district court’s authority over an

offender [must] come[] to an end.” Johnson Br. 31

n.24.

But it is utterly implausible to read the Court’s

opinion as issuing Congress carte blanche to insert

crimes into Section 3583, smoothing the path to

easy—but unreliable and unjust—convictions, by

having them occur in “revocation” proceedings. On the

contrary, as the court of appeals correctly recognized,

the kinds of “serious constitutional problems” that

inhere in pursuing punishment for post-conviction

conduct through informal revocation hearings are the

ones that condemn subsection (k).

III. The Government’s Remedial Proposal Should be

Rejected

The Government finally posits that, accepting

that subsection (k) is unconstitutional, the court of

appeals erred in holding it “facially” so, Br. 51, and

this Court, “at a minimum,” id., should allow the

regime to persist, albeit with juries participating in

the revocation proceedings.

The Government is wrong. The subsection (k)

regime is facially unconstitutional, as that term is

ordinarily understood—indeed on its most restrictive

understanding: There is no set of defendants to whom

the regime Congress enacted may be constitutionally

applied. See United States v. Salerno, 481 U.S. 739,

745 (1987). And the Government’s proposed “remedy”

is neither constitutionally sufficient nor proper.

The constitutional defect of the subsection (k)

regime is that it brings the full, maximal power of the

criminal law to bear in a proceeding that lacks any of

the incidents and protections of a criminal trial, a

proceeding in fact designed for entirely different,

nonpunitive supervisory purposes. The only plainly

adequate remedy for that violation one that: requires

prosecutors to pursue persons believed to have

committed offenses under the enumerated chapters of

Title 18 in criminal trials; forbids judges from

themselves initiating (and then presiding over)

proceedings where defendants face life sentences,

ostensibly for breaching the court’s trust; and ensures

that defendants will not be threatened with a second

“independent prosecution,” Br. 48, at which the

Government pursued further punishment for the

same federal offense. Cf. Dixon, 509 U.S. at 697

(Opinion of Scalia, J.).

That the violations respondent’s Fifth and Sixth

Amendment were manifest and sufficient to overturn

his sentence is no basis for assuming that the

proposed replacement—denying the many historic

safeguards—would comply with the Constitution, a

requirement for any “remedy” a court could impose.

The interests in quick, flexible, and informal hearings

that occur under the very different and far less

draconian subsection (e)(3) revocation regime, are

largely inapplicable and adequate to deny full

protections where life-long liberty deprivation is on

the table. And cases like this one, where such weighty

individual interests are at stake are not the proper

forum for conducting experiments in how little

process may be due.

Moreover, the hybrid proceeding the Government

invites the Court to launch would be an untested

judicial experiment. The statutory scheme does not

establish that Congress would readily replace the

benefits of a unitary, holistic revocation process, with

an as-yet undetermined bi-furcated proceeding, in

order to pursue the subsection (k) punishments. The

remedy the court of appeals adopted, in contrast,

enables the Government both to punish and deter

repeat offenses by those on supervised release the

same way it addresses re-offense by persons who have

regained absolute liberty and those who have not yet

attained release—and to address breaches of trust

through the same subsection (e)(3) mechanism

Congress adjudged adequate for undeniably serious

violations.

Conclusion

The judgment of the court of appeals should be

affirmed.

Respectfully submitted,